Research in physical medicine and rehabilitation. III. The chart review or how to use clinical data for exploratory retrospective studies.

This article guides the design and execution of a small research project using existing clinical data. The most important point in experimental design is to identify potential difficulties and limitations before you start by 1) review of published studies, 2) use of your clinical experience and 3) review of individual case records. Some of these can be addressed by changes in the study design, but some are inherent in the data and methods we are forced to use. The choice of study design may be based primarily on the quality of the clinical data and available resources for additional data collection. The level of measurement (nominal, ordinal, interval or ratio) of your data must first be determined as it limits the descriptive and statistical techniques you can use. After you decide how many variables to include, a rough guess of sample size will help you select your charts for review. Actual review of three charts will further pinpoint any difficulties and will allow you to revise your study and make an accurate estimate of time to completion. Given the long time span of most projects, accurate record keeping is essential.